algorithmic discrimination the challenge of unveiling
TRANSCRIPT
MARCELA MATTIUZZO
Algorithmic Discrimination – The Challenge of Unveiling
Inequality in Brazil
Masters Dissertation
Supervisor: Professor Virgílio Afonso da Silva
UNIVERSITY OF SÃO PAULO
FACULTY OF LAW
São Paulo
2019
MARCELA MATTIUZZO
Algorithmic Discrimination – The Challenge of Unveiling
Inequality in Brazil
Masters dissertation submitted to the graduate
program in law, at the School of Law at the
University of São Paulo, within the field of
concentration of Constitutional Law, under the
supervision of Professor Virgílio Afonso da
Silva.
UNIVERSITY OF SÃO PAULO
SCHOOL OF LAW
São Paulo
2019
Catalogação na Publicação
Serviço de Processos Técnicos da Biblioteca da
Faculdade de Direito da Universidade de São Paulo
Mattiuzzo, Marcela
Algorithmic Discrimination - The Challenge of Unveiling
Inequality in Brazil / Marcela Mattiuzzo. -- São Paulo, 2019.
145 p. ; 30 cm.
Dissertação (Mestrado) – Programa de Pós-Graduação em
Direito, Faculdade de Direito, Universidade de São Paulo, São Paulo,
2019.
Orientador: Virgílio Afonso da Silva.
1. Discriminação algorítmica. 2. Inteligência artificial. 3.
Igualdade. 4. Big Data. 5. Governança algorítmica. I. Silva, Virgílio
Afonso, orient. II. Título.
MATTIUZZO, M. Algorithmic Discrimination – The Challenge of Unveiling Inequality in
Brazil. 2019. 145 f. Dissertação (Mestrado em Direito Constitucional) – Faculdade de Direito,
Universidade de São Paulo, São Paulo, 2019.
Aprovado em:
Banca Examinadora
Prof(a). Dr(a). ____________________________________________
Instituição: ____________________________________________
Julgamento: ____________________________________________
Prof(a). Dr(a). ____________________________________________
Instituição: ____________________________________________
Julgamento: ____________________________________________
Prof(a). Dr(a). ____________________________________________
Instituição: ____________________________________________
Julgamento: ____________________________________________
Acknowledgments
I would like to thank my adviser, Professor Virgílio Afonso da Silva, who indeed made
this dissertation much better than it initially was. And also my “informal advisers,” Artur
Péricles Lima Monteiro, Flávio Marques Prol, Laura Schertel Mendes, and Victor Oliveira
Fernandes, for your many inputs.
I am grateful to my family and to my friends, who provided the most important kind of
assistance: encouragement. And to everyone at VMCA, especially Vinicius Marques de
Carvalho, who worked extra hours to allow me to finish this work.
Quem sabe direito o que uma pessoa é? Antes sendo: julgamento é sempre defeituoso, porque
o que a gente julga é o passado.
Guimarães Rosa
Abstract
MATTIUZZO, Marcela. Algorithmic Discrimination – The Challenge of Unveiling Inequality
in Brazil. 2019. p. 145. Dissertation (Master Degree in Constitutional Law) – School of Law,
University of São Paulo, São Paulo, 2019.
Abstract: the objective of this work is to provide some clarity on what the role of the Law can
be in shedding light upon algorithmic discrimination, as well as how legal instruments could
help minimize its risks, with a specific focus on the Brazilian jurisdiction.
To do so, it first engages in a debate about what algorithms indeed are, and how the emergence
of the data-driven economy, Big Data, and machine learning have leveraged the use of
automated systems. Next, it conceptualizes discrimination, and suggesting a typology of
algorithmic discrimination that takes statistics into account to provide a rationalization of the
debate.
It moves on to discussing the path towards enforcing legal norms against discriminatory
outcomes running from the use of algorithms. Because legislation specifically aimed at fighting
automated systems is still scarce (or application of the current legislation to the problem is
contentious), it engages in a debate about the horizontal effects of fundamental rights – given
that a relevant part of discriminatory practices occur among private parties, and the most basic
defense an individual has against discrimination is the constitutional right to equality. It then
analyzes ordinary legislation in three jurisdictions, the United States of America, Germany, and
Brazil, that could also be enforced against discriminatory practices running from algorithms,
with a special focus on the Brazilian legislation. The legislative debate concludes with the
presentation of two concrete cases of algorithmic discrimination, one concerning the
unemployment policy in Poland, and the other regarding credit scoring in Brazil. The cases are
presented so that the applicability of Brazilian legislation to deal with algorithmic discrimination
can be discussed.
The final chapter is focused on debating the path forward and what can and should be done by
experts, legislators, and policymakers to foster algorithmic innovation without losing sight of
its potential for discrimination. It first presents the literature on algorithmic governance and the
many proposals for dealing with the problem – dedicating a specific section to the challenges
brought about by machine learning – and then sets out an agenda for Brazil.
Keywords: algorithmic discrimination, artificial intelligence, equality, big data, algorithmic
governance.
Resumo
MATTIUZZO, Marcela. Discriminação Algorítmica – O desafio em desvendar a desigualdade
no Brasil. 2019. p. 145. Dissertação (Mestrado em Direito Constitucional) – Faculdade de
Direito, Universidade de São Paulo, São Paulo, 2019.
Resumo: O objetivo desse trabalho é esclarecer qual é o papel que o Direito pode desempenhar
no debate sobre a discriminação algorítmica, assim como de que maneira os instrumentos
jurídicos podem auxiliar a mitigar os riscos discriminatórios desse tipo de prática, com foco
especial na jurisdição brasileira.
Para isso, primeiro o trabalho propõe um debate sobre o que são algoritmos, e como a
emergência da economia de dados, do Big Data e de técnicas de machine learning impulsionam
o uso de sistemas automatizados. Em seguida, conceitua-se a discriminação, propondo-se uma
tipologia para a discriminação algorítmica que leva em conta questões estatísticas, a fim de
racionalizar a discussão.
A dissertação então parte para o debate sobre os caminhos para a aplicação de normas jurídicas
em face de discriminação algorítmica. Dado que leis e normas especificamente voltados a esse
tema ainda não são muito difundidas (e que a aplicação da legislação existente a essa questão é
controversa), o trabalho propõe um debate sobre a eficácia horizontal dos direitos fundamentais
– tendo em vista que boa parte das práticas discriminatórias via uso de algoritmos se dá entre
partes privadas, e que a defesa mais básica que um indivíduo tem contra a discriminação é o
direito constitucionalmente garantido à igualdade. Passa-se então a uma análise da legislação
ordinária em três jurisdições, Estados Unidos da América, Alemanha e Brasil, legislação essa
que pode também ser aplicada em casos de práticas discriminatórias levadas a cabo via
algoritmos, dando especial destaque ao caso brasileiro. Esse debate legislativo é concluído com
a apresentação de dois casos concretos, um que diz respeito à política de acesso a emprego na
Polônia e outro que trata das práticas de credit scoring no Brasil. Os casos são apresentados de
forma a se pensar a eventual possibilidade de uso de regras brasileiras para lidar com os temas
discriminatórios que se colocam concretamente.
O capítulo final tem como foco o debate do caminho a ser trilhado, e qual pode e deve ser feito
por especialistas, legisladores e aplicadores do direito para promover a inovação no campo
algorítmico sem perder de vista seus potenciais impactos discriminatórios. Primeiro, apresenta-
se a literatura sobre governança algorítmica e as muitas propostas que pretendem endereçar o
tema – com especial atenção aos desafios apresentados pelo machine learning – e então delineia-
se uma agenda para o Brasil sobre o assunto.
Palavras-chave: discriminação algorítmica, inteligência artificial, igualdade, big data,
governança algorítmica.
Table of Contents
1 Introduction ............................................................................................................................ 17
2 Algorithms, Models, and Discrimination .............................................................................. 25
2.1 The Algorithmic System ................................................................................................. 25
2.2 The Emergence of Big Data and Machine Learning ...................................................... 27
2.2.1 The risk of discrimination ........................................................................................ 33
2.3 Discrimination and Profiling - Generalizations under the law ....................................... 39
2.3.1 Striving for Particularization - Is individualized decision-making superior? .......... 39
2.3.2 Profiling and Other Legal Matters ........................................................................... 44
2.3.3 Algorithmic Discrimination: A Proposed Typology ............................................... 46
2.3.4 Causation, Correlation, Objectivity and The Limitations of Algorithms ................ 53
2.3.4.1 Objectivity ........................................................................................................ 56
2.3.4.2 Prediction, not Judgment .................................................................................. 58
3 Algorithmic Discrimination in Practice – The Challenges in Enforcement .......................... 61
3.1 The Horizontal Effects of Fundamental Rights .............................................................. 62
3.1.1 United States of America ......................................................................................... 62
3.1.2 Germany ................................................................................................................... 66
3.1.3 Brazil ........................................................................................................................ 69
3.1.4 Relevance for algorithmic discrimination ................................................................ 73
3.2 Antidiscrimination and Data Protection Legislation – What Lies Beyond Fundamental
Rights .................................................................................................................................... 74
3.2.1 The United States and the Civil Rights Act ............................................................. 74
3.2.2 Germany, informational self-determination, and the European Union ................... 77
3.2.3 Brazil, the Right to Equality, and the GDPA ........................................................... 82
3.2.3.1 The Consumer Protection Code ........................................................................ 82
3.2.3.2 The Credit Information Act ............................................................................... 85
3.2.3.3 The Public Information Access Act .................................................................. 89
3.2.3.4 The Brazilian Internet Framework .................................................................... 90
3.2.3.5 The General Data Protection Act ...................................................................... 90
3.3 Case Studies .................................................................................................................... 93
3.3.1 Labor and unemployment in Poland ........................................................................ 93
3.3.2 Credit scoring in Brazil ............................................................................................ 99
4 Algorithmic Discrimination and the Law – The Way Forward ........................................... 103
4.1 Algorithmic Governance and Policy Proposals – From Transparency to Accountability
............................................................................................................................................. 103
4.1.1 The Policy Challenges of Machine Learning ......................................................... 117
4.2 An Agenda for Brazil .................................................................................................... 125
References ............................................................................................................................... 133
17
1 Introduction
The game of Go was developed in China over 2,500 years ago, it is the oldest board
game that human kind is aware of, and counts with a legion of fans. The objective of the game
is simple: each of the two players must occupate more board territory than the opponent in order
to win, either by capturing opponent’s stones or by surrounding empty space. The black and
white stones are the tools the players must handle to reach their objective; and they both take
turns placing them on the board. Apparently, the game is very simple, but in reality it can be
terribly complex. The reason is the number of possible plays: whereas in chess the initial move
by any player is limited to 20 moves, in Go it spikes to 361, and grows exponentially thereafter.
To have an idea, the estimated number of possible board configurations in chess is 10120,
whereas in Go it is 10174, meaning there are trillions more configurations in Go than in chess.
In 2016, DeepMind, a research project turned start-up, later acquired by Google’s parent
company Alphabet, challenged world-champion Lee Sedol to a game of Go.1 The challenger:
AlphaGo, a deep neural network developed by a group of programmers. Demis Hassabis,
DeepMind’s CEO and co-founder, a child chess prodigy turned computer scientist, entered into
this endeavor after deciding that to test the true potential of the topic he had been studying since
graduation from Cambridge University – artificial intelligence – he had to crack the game of Go
by beating a professional player.
AlphaGo beat Sedol 4-1 in a 5-match challenge. As Hassabis explains in the AlphaGo
documentary, originally the program learned how to play by watching over 100,000 games by
strong amateurs, and its objective was to mimic human players. What gave it the real leap,
however, was learning from its own mistakes. It has since improved its capacity by playing
games against itself, and became the best Go player in the world.2 This celebrated achievement
of artificial intelligence is by no means the only application of algorithms to problem-solving,
1 Before challenging Sedol, AlphaGo had already beat the European Champion, Fan Hui. The results of the
European matches were published by DeepMind in Nature. See: SILVER, D. et al. Mastering the game of Go
with deep neural networks and tree search. Nature, vol. 529, January 28, 2016. Available at:
<https://storage.googleapis.com/deepmind-media/alphago/AlphaGoNaturePaper.pdf>. Access: January 10, 2019. 2 SILVER, D. et al. Mastering the game of Go without human knowledge. Nature, vol. 550, January 19, 2017.
18
quite on the contrary; algorithms have been applied to a wide-array of matters with various
outcomes, notably, they have been used by public authorities to inform decisions on various
matters, ranging from sentencing to determining the best treatment for a given illness.
Imagine that one day, instead of leaving decisions of whether or not an individual should
be imprisoned to a judge, an algorithm3 was employed to calculate the risk of convicts’ future
conduct as the basis for the criminal sanction; that is, the prison term would vary depending on
whether the algorithm identified convicts as presenting “low” or “high” risk to society. Imagine
that to reach its decision the algorithm evaluated the convicts’ responses to a wide array of
questions that included whether the person’s parents had ever been incarcerated, whether her
friends had ever consumed illegal drugs, how often this person had been involved in fights at
school, and so on. Imagine, lastly, that the mathematical formulas and weighting used by this
algorithm were not public, such that there was no public access to the input used or outputs
generated by it, and that convicts were thus unable to determine which aspects of their behavior
led to their classification as high or low risk to society.
This scenario might sound like science fiction, but algorithms such as the one described
above are already in use in several jurisdictions within the United States, and their
implementation has also begun in Brazil.4 Much about their use is questionable,5 but this work
3 One of the goals of this work is to clarify what is meant by an algorithm and how it carries out tasks such as risk
assessment. In short, an algorithm is nothing more than a sequence of actions to be performed in a given order,
which generates a specific result. One can have an algorithm for making dinner, an algorithm for taking a bath, an
algorithm for walking to work every day, etc. Thus, algorithms do not have to be computerized, nor even need to
be complex. 4 The most popular tools are COMPAS and LSI-R (Level of Service Inventory Revised). In Brazil, the only instance
of algorithmic use in sentencing so far is in the State of Minas Gerais, where an algorithm was used to identify the
kinds of appeals presented by parties and to provide “model sentences” for review by judges. Other algorithms,
however, are frequently used for purposes such as selecting Reporting Justices at the Brazilian Supreme Court.
See: Tribunal de Justiça do Estado de Minas Gerais. TJMG utiliza inteligência artificial em julgamento virtual.
November 07, 2018. Available at: <https://www.tjmg.jus.br/portal-tjmg/noticias/tjmg-utiliza-inteligencia-
artificial-em-julgamento-virtual.htm#.XC1Vby2ZO1s>. Access: January 05, 2019. and Supremo Tribunal Federal.
Ministra Carmen Lúcia anuncia início de funcionamento do Projeto Victor, de inteligência artificial. Notícias
STF. August 30, 2018. Available at: <
http://www.stf.jus.br/portal/cms/verNoticiaDetalhe.asp?idConteudo=388443>. Access: January 05, 2019. 5 When it comes to the legality of their adoption, though the law varies from one jurisdiction to the next, democratic
nations are unanimous in affirming the requirement of some form of reasoned motivation for court decisions, and
that rationale cannot be entirely random, or must at least be sufficiently clear to allow the interested party to appeal
it. In Brazil, Article 93, IX of the Constitution says that “all judgments of judicial authorities shall be public, and
all decisions shall be motivated”. In common law jurisdictions, the duty to give reasons may not exist when it
19
primarily focuses n the discriminatory risks of the output generated by algorithmic systems.6
According to Julia Angwin and her team at ProPublica who revealed the use of algorithms in
criminal sentencing and its legal and moral implications, COMPAS, the tool developed by the
company Northpointe (later renamed Equivant) adopted by the justice system in states such as
Florida, “turned up significant racial disparities (…) In forecasting who would re-offend, the
algorithm made mistakes with black and white defendants at roughly the same rate but in very
different ways”. Whereas blacks were falsely flagged as high risk and potential re-offenders at
twice the rate as white defendants, whites were more frequently deemed as low risk than black
defendants.7
Though the use of algorithmic systems for sentencing has yet to reach the same intensity
in Brazil as in the United States, discrimination through automation is already part of our reality
and, given the extensive use of such tools in other jurisdictions, the debate over them will only
comes to administrative decisions, but judicial duty to give reasons is upheld. In: LO, H. The Judicial Duty to
Give reasons. Legal Studies, vol. 20, issue 1, March 2000, p. 42 - 65. 6 Other repercussions include the risks to data privacy and security, cf. section 2.3.2. There is a natural connection
between the two, but in protecting individuals from discrimination we may not always be ensuring greater privacy. 7 ANGWIN, Julia; LARSON, Jeff; MATTU, Surya; KIRCHNER, Lauren. Machine Bias. ProPublica, May 23,
2016. Available at: <https://www.propublica.org/article/machine-bias-risk-assessments-in-criminal-sentencing>.
Access: January 10, 2019., in which they state: “We also turned up significant racial disparities, just as Holder
feared. In forecasting who would re-offend, the algorithm made mistakes with black and white defendants at
roughly the same rate but in very different ways. The formula was particularly likely to falsely flag black defendants
as future criminals, wrongly labeling them this way at almost twice the rate as white defendants. White defendants
were mislabeled as low risk more often than black defendants. Could this disparity be explained by defendants’
prior crimes or the type of crimes they were arrested for? No. We ran a statistical test that isolated the effect of race
from criminal history and recidivism, as well as from defendants’ age and gender. Black defendants were still 77
percent more likely to be pegged as at higher risk of committing a future violent crime and 45 percent more likely
to be predicted to commit a future crime of any kind.”
20
grow. Algorithms have been deployed for making diagnoses,8 job selection,9 and for even
welfare eligibility processes.10
The present work surveys the consequences of algorithmic discrimination, as well as the
paths for the protection against discriminatory conduct. Four important observations must be
made in this introduction: first, as will be further scrutinized throughout this dissertation,
discrimination is not the only issue that may arise from algorithmic use. Problems such as
privacy infractions may also run from algorithmic use but do not necessarily relate to
discrimination, as section 2.3.2 will demonstrate. These problems are not within the scope of
my work.
Second, if automation and its potential for discrimination are cause for concern, we must
also remember that the world is full of generalizations, and that decision-making largely
depends on generalizing to be feasible. The legal system, in particular, is full of necessary
generalizations. Whenever we set out a rule, for example one that makes it illegal to drive above
80 km/h on a given road, or when we establish that only those over a given age have the right
to vote, we are using, engaging, and reinforcing generalizations. Society as a whole accepts this
manner of decision and rule-making, which leads us to suspect that the problem with profiling
8 “Complex algorithms will soon help clinicians make incredibly accurate determinations about our health from
large amounts of information, premised on largely unexplainable correlations in that data. […] With extraordinary
accuracy, these algorithms were able to predict and diagnose diseases, from cardiovascular illnesses to cancer, and
predict related things such as the likelihood of death, the length of hospital stay, and the chance of hospital
readmission. Within 24 hours of a patient’s hospitalization, for example, the algorithms were able to predict
with over 90% accuracy the patient’s odds of dying. These predictions, however, were based on patterns in the data
that the researchers could not fully explain.” In: BURT, A. and VOLCHENBOUM, S. How Health Care Changes
When Algorithms Start Making Diagnoses. Harvard Business Review, May 08, 2018. Available at:
<https://hbr.org/2018/05/how-health-care-changes-when-algorithms-start-making-diagnoses>. Access: January
10, 2019. 9 “Many companies, including Vodafone and Intel, use a video-interview service called HireVue. Candidates are
quizzed while an artificial-intelligence (AI) program analyses their facial expressions (maintaining eye contact with
the camera is advisable) and language patterns (sounding confident is the trick). People who wave their arms about
or slouch in their seat are likely to fail. Only if they pass that test will the applicants meet some humans.” THE
ECONOMIST. How an algorithm may decide your career. June 21, 2018. Available at:
<https://www.economist.com/business/2018/06/21/how-an-algorithm-may-decide-your-career>. Access: January
10, 2019. 10 Several examples of welfare automation in the United States are investigated in: EUBANKS, V. Automating
Inequality: How High-Tech Tools Profile, Police, and Punish de Poor. St. Martin’s Press. January 23, 2018.
Eubanks provides examples on how welfare eligibility in Indiana, the homeless policy in Los Angeles, and others
were automated.
21
and discrimination is not the practice of making generalizations in itself, but rather with the
criteria used to create groups that are allowed (or not) to behave a certain way or access a specific
good. Therefore, this dissertation takes the view, further scrutinized in section xxx, that
discrimination is not always harmful, nor is it always illegal.
Third, the advent of machine learning has brought a number of complexities to this
landscape. A system such as AlphaGo, which works with deep neural networks, is very different
from a system like COMPAS, whose functioning is based on “old-fashioned” programming.
The consequences for discrimination are severe, especially when it comes to identifying
solutions. As section 4.1.1 further investigates, asking for transparency of machine learning
algorithms is of little use, and accountability proposals are also challenging to implement.
Lastly, it should be clear from the start that this dissertation does not in any way suggest
that innovation should be contained, for it always leads to unwanted outcomes. It merely intends
to better understand and investigate one potentially harmful outcome, which is quite different.
As with most of technology that came before algorithms, implementation always presents pros
and cons; the challenge for society is to learn how to strike a balance that allows for the benefits
and contains the disadvantages.
For no other reason the discussion is increasingly present in the private sector and also
among public authorities. In March 2018, the Council of Europe, through its Committee of
experts on Internet Intermediaries, finalized and published a study on the human rights
dimensions of automated data processing techniques, focusing primarily on algorithms and
possible regulatory implications.11 The objective of the study is to
map out some of the main current concern from the Council of Europe’s human
rights perspective, and to look at possible regulatory options that member
states may consider to minimise adverse effects, or to promote good
practices.12
11 COUNCIL OF EUROPE PORTAL. Algorithms and Human Rights: a new study has been published. March
22, 2018. Available at: <https://www.coe.int/en/web/freedom-expression/-/algorithms-and-human-rights-a-new-
study-has-been-published>. Access> January 10, 2019. 12 Ibid, p. 4.
22
Similarly, in May 2018, the Science and Technology Committee of the House of
Commons in the United Kingdom published a report on decision-making by algorithms, in order
to “identify the themes and challenges that the proposed Center for Data Ethics & Innovation
[which shall be created by the British government shortly13] should address as it begins its
work”. The report calls attention to the challenges brought about by automated decisions,
especially bias, as well as the possible solutions for such hurdles, mainly involving mechanisms
for transparency and accountability.14
The United States is also taking part in the debate. In 2016, the Federal Trade
Commission issued a report on the consequences of Big Data and its inclusionary and
exclusionary uses.15 The FTC states that
Big data analytics can provide numerous opportunities for improvements in
society. In addition to more effectively matching products and services to
consumers, big data can create opportunities for low-income and underserved
communities. […] At the same time, workshop participants and other have
noted how potential inaccuracies and biases might lead to detrimental effects
for low-income and underserved populations. For example, participants raised
concerns that companies could use big data to exclude low-income and
underserved communities from credit and employment opportunities.16
Australia was one of the first countries to expressly address the issues brought about by
big data, automation, and algorithms. Back in 2007, the Australian government had already
issued its Better Practice Guide for automated decision-making in the public administration. The
Guide is an effort to consolidate and provide concrete solutions for the best practice principles
of the Automated Assistance in Administrative Decision-Making Report issued by the Attorney-
General in 2004. It states that “[a]utomated systems have been used for some users in areas of
13 DEPARTMENT FOR DIGITAL, CULTURE, MEDIA & SPORT of the UK Government. Consultation on the
Centre for Data Ethics and Innovation. November 20, 2018. Available at:
<https://www.gov.uk/government/consultations/consultation-on-the-centre-for-data-ethics-and-innovation/centre-
for-data-ethics-and-innovation-consultation>. Access: January 10, 2019. 14 SCIENCE AND TECHNOLOGY COMMITTEE of the House of Commons. Algorithms in decision-making.
Fourth Report of Session 2017-19. May 23, 2018. Available at:
<https://publications.parliament.uk/pa/cm201719/cmselect/cmsctech/351/351.pdf>. 15 FEDERAL TRADE COMMISSION. Big Data, A Tool for Inclusion or Exclusion? – Understanding the
Issues. January, 2016, p. 1. Available at: <https://www.ftc.gov/system/files/documents/reports/big-data-tool-
inclusion-or-exclusion-understanding-issues/160106big-data-rpt.pdf>. Access: January 01, 2019. 16 Ibid, p. i.
23
financial entitlement for citizens and other agency customer groups, such as welfare, family and
veteran support benefits.”17
Brazil, on its part, has so far been less explicit in this debate. The Brazilian Strategy para
Digital Transformation issued in 2018 recognizes the relevance of the data-driven economy and
sets forth strategies the government should pursue to extract value from this new environment,
namely: (i) promote the approval of incentives to attract data centers to Brazil; (ii) improve the
National Policy on Government Open Data, and foster the adoption of tools, systems, and
processes based on data; (iii) promote cooperation among authorities and the harmonization of
normative frameworks regarding data, in order to facilitate the inclusion of Brazilian companies
in the global market; (iv) promote cooperation among government representatives, universities,
and companies, as to facilitate the knowledge and technology exchange; (v) stimulate the
adoption of cloud services as part of the public administration’s services and data storage
systems; and (vi) evaluate the potential social and economic effects of disruptive technologies,
such as artificial intelligence and big data, proposing policies that aim at mitigating their
negative effects and simultaneously maximize their positive effects.18
It is in this context that this work aims at answering the following question: What ought
be the role of the Law in illuminating and addressing algorithmic discrimination in the Brazilian
context?
Chapter 2 provides the basis upon which this work will be built. It explains in further
detail how the emergence of Big Data was crucial in allowing algorithms to evolve. It also
highlights what algorithmic discrimination is, and what it is not. It is in this chapter that I
propose a typology for algorithmic discrimination, with the goal of contributing to the
discussion and hopefully helping bring more clarity to the topic. Chapter 3 delineates the legal
framework in which algorithmic discrimination will be discussed. It presents the debate in
17 AUSTRALIAN GOVERNMENT. Automated Assistance in Administrative Decision-Making – Better
Practice Guide. February, 2007. Available at:
<https://www.oaic.gov.au/images/documents/migrated/migrated/betterpracticeguide.pdf>. Access: January 01,
2019. 18 MCTIC. Estratégia Brasileira para a Transformação Digital (E-Digital). Brasília, 2018, p. 64-65. Available
at: <http://www.mctic.gov.br/mctic/export/sites/institucional/estrategiadigital.pdf>. Access: January 12, 2019.
24
constitutional terms, and also in terms of ordinary legislation, focusing primarily (though not
solely) in the case of Brazil. It applies the enforcement debate to two concrete cases, one which
took place in Poland, regarding unemployment public policy, and the other in Brazil, regarding
credit scoring. Chapter 4 is focused on debating policy solutions, by presenting a review of the
literature on algorithmic governance as it currently stands, and discussing an agenda for the
Brazilian jurisdiction.
133
References
ACHARYA, V. Regulating wall Street: the Dodd-Frank Act and the new architecture of
global finance. Interviewer: DAVIES, V. VOX - CEPR Policy Portal. October 22, 2010.
Available at: <https://voxeu.org/vox-talks/regulating-wall-street-dodd-frank-act-and-new-
architecture-global-finance>.
AGRAWAL, A., GANS, J. and GOLDFARB, A. Prediction Machines: The Simple
Economics of Artificial Intelligence. Harvard Business Press, April 17, 2018.
ANANNY, M. and CRAWFORD, K. Seeing without knowing: Limitations of the
transparency ideal and its application to algorithmic accountability. SAGE journals.
December 13, 2016. Available at:
<https://journals.sagepub.com/doi/abs/10.1177/1461444816676645?journalCode=nmsa>.
ANBC. Por que a ANBC?. Available at:
<https://www.anbc.org.br/materias.php?cd_secao=3#5&friurl=_-Empresas-associadas-_>.
ANDERSON, C. The End of the Theory: the Data Deluge Makes the Scientific Method
Obsolete. Wired. June 23, 2008. Available at: <https://www.wired.com/2008/06/pb-theory/>.
ANGWIN, Julia; LARSON, Jeff; MATTU, Surya; KIRCHNER, Lauren. Machine Bias.
ProPublica, May 23, 2016. Available at: <https://www.propublica.org/article/machine-bias-
risk-assessments-in-criminal-sentencing>.
ASSOCIATION FOR COMPUTING MACHINERY US PUBLIC POLICY COUNCIL
(USACM). Statement on Algorithmic Transparency and Accountability. January 12, 2017.
Available at: <http://www.acm.org/binaries/content/assets/public-
policy/2017_usacm_statement_algorithms.pdf>.
AUSTRALIAN GOVERNMENT. Automated Assistance in Administrative Decision-
Making – Better Practice Guide. February, 2007. Available at:
<https://www.oaic.gov.au/images/documents/migrated/migrated/betterpracticeguide.pdf>.
BANERJEE, A. DUFLO, E. GLENNERSTER, R. and KINNAN, C. The miracle of
microfinance? Evidence from a randomized evaluation. Northwestern University of
Economics and NBER. March, 2014. Available at: <https://economics.mit.edu/files/5993>.
BAROCAS, S. and SELBST, A. D. Big Data’s Disparate Impact. p. California Law Review,
vol. 671, 2016. Available at: <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2477899>.
BERGER, M. M. To Regulate, or Not to Regulate – Is That the Question: Reflections on
the Supposed Dilemma between Environmental Protection and Private Property Rights.
134
8 Loy, L. A. L. Rev. 253, 1975. Available at:
<https://digitalcommons.lmu.edu/cgi/viewcontent.cgi?referer=https://scholar.google.com.br/sc
holar?hl=en&as_sdt=0%2C5&q=to+regulate+or+not&btnG=&httpsredir=1&article=1187&co
ntext=llr>.
BILBAO UBILLOS, J. M. La Eficacia de Los Derechos Fundamentales Frente a
Particulares: Analisis de La Jurisprudencia del Tribunal Constitucional. Centro de
Estudios Políticos y Constitucionales, 1997.
BLASS A. and GUREVICH, Y.. Algorithms: A Quest for Absolute Definitions. Bulletin of
European Association for Theoretical Computer Science. vol. 81, 2003. Available at:
<https://web.eecs.umich.edu/~gurevich/Opera/164.pdf>.
BOYD, D. and CRAWFORD, K. Six Provocations for Big Data. A Decade in Internet Time:
Symposium on the Dynamics of the Internet and Society, September 2011. Available at:
<https://papers.ssrn.com/sol3/papers.cfm?abstract_id=1926431>.
BRILL, C. J. Scalable Approaches to Transparency and Accountability in Decision Making
Algorithms, Remarks at the NYU Conference on Algorithms and Accountability. FTC.
February 28, 2015. Available at: <https://www.ftc.gov/public-statements/2015/02/scalable-
approaches-transparency-accountability-decisionmaking-algorithms>.
BULLUNCK, M. Histories of algorithms: Past, present and future. Historia Mathematica,
Elsevier, 2015, 43 (3).
CENTRO DE ESTUDOS DA MTERÓLE (CEM). Mapa da Vulnerabilidade Social da
População da Cidade de São Paulo. Centro Brasileiro de Análise e Planejamento-CEBRAP,
do Serviço Social do Comércio-SFSC e da Secretaria Municipal de Assistência Social de São
Paulo, SAS-PMSP. São Paulo, 2004. Available at:
<http://web.fflch.usp.br/centrodametropole/upload/arquivos/Mapa_da_Vulnerabilidade_social
_da_pop_da_cidade_de_Sao_Paulo_2004.pdf>.
CITRON, D. K. Citron, Technological Due Process, 85 Wash. U. L. Rev. 1249, 2008.
Available at: <http://openscholarship.wustl.edu/law_lawreview/vol85/iss6/2>.
CITRON, D. K. and PASQUALE, F. The Scored Society: Due Process For Automated
Predictions. Washington Law Review, vol. 89:1, 2014. Available at:
<https://digital.law.washington.edu/dspace-
law/bitstream/handle/1773.1/1318/89WLR0001.pdf>.
COMMISSION OF THE EUROPEAN COMMUNITIES. Council Directive on implementing
the principle of equal treatment between persons irrespective of religion or belief,
disability, age or sexual orientation. Brussels, 2008. Available at: <https://eur-
lex.europa.eu/legal-content/en/TXT/?uri=CELEX%3A52008PC0426>.
135
COMMUNICATIONS COMMITTEE. The Internet: to regulate or not to regulate? inquiry.
Parliamentary business. Available at:
<https://www.parliament.uk/business/committees/committees-a-z/lords-
select/communications-committee/inquiries/parliament-2017/the-internet-to-regulate-or-not-
to-regulate/>.
CORMEN, T. H., Algorithms Unlocked. MIT Press, 2013.
COUNCIL OF EUROPE PORTAL. Algorithms and Human Rights: a new study has been
published. May 22, 2018. Available at: <https://www.coe.int/en/web/freedom-expression/-
/algorithms-and-human-rights-a-new-study-has-been-published>.
CRAWFORD, H. The Hidden Biases in Big Data. Harvard Business Review. April 01, 2013.
Available at: <https://hbr.org/2013/04/the-hidden-biases-in-big-data>.
CUMMINGS, M. L. The Social and Ethical Impact of Decision Support Interface Design.
International Encyclopedia of Ergonomics and Human Factors. 2005. Available at:
<https://pdfs.semanticscholar.org/a9b3/ec436508ebfa40f3a3f5b59231bece4f3e34.pdf>.
DEMSETZ, H. Why Regulate Utilities?. Journal of Law and Economics, vol. 11, no. 1, The
University of Chicago Press Journals, April, 1968. Available at:
<https://www.journals.uchicago.edu/doi/abs/10.1086/466643?journalCode=jle>.
DEPARTMENT FOR DIGITAL, CULTURE, MEDIA & SPORT of the UK Government.
Consultation on the Centre for Data Ethics and Innovation. November 20, 2018. Available
at: <https://www.gov.uk/government/consultations/consultation-on-the-centre-for-data-ethics-
and-innovation/centre-for-data-ethics-and-innovation-consultation>.
DIAKOPOULOS, N. and FIEDLER, S. How to Hold Algorithms Accountable. MIT
Technology Review. November, 2016.
DIAKOPOULOS, N., FIEDLER, S., ARENAS, M. et al. Principles for Accountable
Algorithms and a Social Impact Statement for Algorithms. FAT/ML. Available at:
<https://www.fatml.org/resources/principles-for-accountable-algorithms>.
DOMINGOS, P. Master Algorithm. Basic Books Inc. New York, 2018.
DONEDA, D. and MENDES, L. S. Data Protection in Brazil: New Developments and
Current Challenges. In: GURWIRTH, S., LEENES, R. and HERT, P. De. (Eds). Reloading
Data Protection: Multidisciplinary Insights and Contemporary Challenges. Springer, 2014.
DOSHI-VELEZ, F. and KORTZ, M. Accountability of AI Under the Law: The Role of
Explanation. Berkman Klein Center Working Group on Explanation and the Law, Berkman
136
Klein Center for Internet & Society working paper, 2017. Available at :
<https://dash.harvard.edu/bitstream/handle/1/34372584/2017-11_aiexplainability-
1.pdf?sequence=3>.
DÜRIG, G. Grundrechte und Zivilrechtsprechung. Vom Bonner Grundgesetz zur
gesamtdeutschen Verfassung : Festschrift zum 75. Geburtstag von Hans Nawiasky, 1956.
EDWARDS, L. and VEALE, L. Slave to the Algorithm? Why a ‘Right to an Explanation’
Is Probably Not the Remedy You Are Looking For. 16 Duke Law & Technology Review 18,
2017. Available at:
<https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2972855&download=yes>.
ELLIS, E. and WATSON, P. EU Anti-Discrimination Law - Second Edition. Oxford EU Law
Library, 2012.
ELLROTT, J. TRITTMANN, R. and WERMEISTER, C. Automated driving law passed in
Germany. June 21, 2017. Available at: <https://www.freshfields.com/en-gb/our-
thinking/campaigns/digital/internet-of-things/connected-cars/automated-driving-law-passed-
in-germany/>.
EUBANKS, V. Automating Inequality: How High-Tech Tools Profile, Police, and Punish
de Poor. St. Martin’s Press. January 23, 2018.
EUROPEAN COMMISSION. Charter of Fundamental Rights: the Presidents of the
Commission, European Parliament and Council sign and solemnly proclaim the Charter
in Strasbourg. Brussels, December 12, 2007. Available at: <http://europa.eu/rapid/press-
release_IP-07-1916_en.htm>.
EUROPEAN COMMISSION’S HIGH-LEVEL EXPERT GROUP ON ARTIFICIAL
INTELLIGENCE (AI HELG). Draft Ethics Guidelines for Trustworthy AI. European
Commission, Brussels, March, 2019 (final version). Available at:
<https://ec.europa.eu/futurium/en/node/6044>.
FEDERAL TRADE COMMISSION. Big Data, A Tool for Inclusion or Exclusion? –
Understanding the Issues. January, 2016. Available at:
<https://www.ftc.gov/system/files/documents/reports/big-data-tool-inclusion-or-exclusion-
understanding-issues/160106big-data-rpt.pdf>.
FUSTER, G. G. The Emergence of Personal Data Protection as a Fundamental Right of
the EU. Springer International Publishing, 2014.
GARDBAUM, S. The ‘Horizontal Effect’ of Constitutional Rights. Michigan Law Review,
vol. 102, UCLA School of Law Research Paper No. 03-14, pp. 388-459, 2003. Available at:
<https://papers.ssrn.com/sol3/papers.cfm?abstract_id=437440>.
137
GASSER, U. and ALMEIDA, V. A.F. A Layered Model for AI Governance. IEEE Internet
Computing 21 (6) (November): 2017.
GILLESPIE, T. Chapter 2- Algorithm. In: PETERS, B. (Ed.). Digital Keywords: a Vocabulary
of information society and culture. Princeton: Princeton University Press, 2016.
GOODMAN, B. W. A Step Towards Accountable Algorithms?: Algorithmic
Discrimination and the European Union General Data Protection. 29th Conference on
Neural Information Processing Systems. Barcelona, Spain, 2016.
GORZONI, P. F. A. da C. Supremo Tribunal Federal e a Vinculação dos Direitos
Fundamentais nas Relações entre Particulares. 2007. Available at:
<http://www.sbdp.org.br/publication/supremo-tribunal-federal-e-a-vinculacao-dos-direitos-
fundamentais-nas-relacoes-entre-particulares/>.
GUNNING, D. Explained Artificial Intelligence (XAI). DARPA/I20. November, 2017.
Available at: <https://www.darpa.mil/attachments/XAIProgramUpdate.pdf>.
GUREVICH, Y. What Is an Algorithm?. In: BIELIKOVÁ M., FRIEDRICH, G., GOTTLOB,
G., KATZENBEISSER, S., TURÁN, G. (eds) SOFSEM 2012: Theory and Practice of
Computer Science. SOFSEM 2012. Lecture Notes in Computer Science, vol. 7147. Springer,
Berlin, Heidelberg, 2012. Available at:
<https://www.researchgate.net/publication/221512843_What_Is_an_Algorithm>.
HARFORD, T. Big data: are we making a big mistake?. The Financial Times. 2014.
Available at: <https://rss.onlinelibrary.wiley.com/doi/pdf/10.1111/j.1740-9713.2014.00778.x>.
HAWKINS, J. and BLAKESLEE, S. On Intelligence: How a New Understanding the Brain
Will Lead to the of the Creation of Truly Intelligent Machines. Times Books, 2004.
HESSE, K. Verfassungsrecht und Privatrecht. Müller Jur. Vlg. C. F., 1988.
HUET, E. Server And Protect: Predictive Policing Firm PredPol Promises To Map Crime
Before It Happens. Forbes, February 11, 2015. Available at:
<https://www.forbes.com/sites/ellenhuet/2015/02/11/predpol-predictive-
policing/#6efe33ae4f9b>.
IAPP and THE UNITED NATIONS GLOBAL PULSE. Building Ethics into Privacy
Frameworks for Big Data and AI. Available at:
<https://iapp.org/media/pdf/resource_center/BUILDING-ETHICS-INTO-PRIVACY-
FRAMEWORKS-FOR-BIG-DATA-AND-AI-UN-Global-Pulse-IAPP.pdf>.
138
ITS. Transparência e governança nos algoritmos: um estudo de caso sobre o setor de birôs
de crédito. Rio de Janeiro, 2017.
JIA, J., JIN, G. J. and WAGMAN, L. The Short-Run Effects of GDPR On Technology
Venture Investment. NBER Working Paper no. 25248, November, 2018. Available at:
<https://www.nber.org/papers/w25248>.
JUSTICE AND CONSUMERS. Guidelines on Automated Individual decision-making and
Profiling for the purposes of Regulation 2016/679. Adopted on October 03, 2017. Available
at: <https://ec.europa.eu/newsroom/article29/item-detail.cfm?item_id=612053>.
KELLSTEDT, P. M. and WHITTEN, G. The Fundamentals of Political Research. New York:
Cambridge University Press, 2009.
KLEINBERG, J. and MULLAINATHAN, S. Simplicity Creates Inequity: Implications for
Fairness, Stereotypes, and Interpretability. Cornell University, September 12, 2018.
Available at: <https://arxiv.org/abs/1809.04578>.
KNIGHT, W. The Dark Secret at the Heart of AI. MIT Technology Review. April 11, 2017.
Available at: <https://www.technologyreview.com/s/604087/the-dark-secret-at-the-heart-of-
ai/>.
KROLL, J. et al. Accountable Algorithms. 165 U. PA. L. Rev. 633, 2017. Available at:
<https://scholarship.law.upenn.edu/penn_law_review/vol165/iss3/3/>.
LAZER, D. et. al. The Parable of Google Flu: Traps in Big Data Analysis. Science 343:1203-
1205, March 14, 2014. Available at:
<https://gking.harvard.edu/files/gking/files/0314policyforumff.pdf>.
LEMOS, R. et al. A criação da Autoridade Nacional de Proteção de Dados pela MP nº
869/2018. JOTA, December 29, 2018. Available at:
<https://www.jota.info/?pagename=paywall&redirect_to=//www.jota.info/opiniao-e-
analise/artigos/a-criacao-da-autoridade-nacional-de-protecao-de-dados-pela-mp-no-869-2018-
29122018>.
LERMAN, J. Big Data and Its Exclusions. Stanford Law Review Online, 2013.
LESSIG, L. Against Transparency. The New Republic, 2009. Available at:
<https://newrepublic.com/article/70097/against-transparency>.
LEVIN, S. New AI can guess whether you're gay or straight from a photograph. The
Guardian. September 08, 2017. Available at:
139
<https://www.theguardian.com/technology/2017/sep/07/new-artificial-intelligence-can-tell-
whether-youre-gay-or-straight-from-a-photograph>.
LI, F.-F. How to Make A.I. That’s Good for People. The New York Times, March 7, 2018.
Available at: <https://www.nytimes.com/2018/03/07/opinion/artificial-intelligence-
human.html>.
LO, H. The Judicial Duty to Give reasons. Legal Studies, vol. 20, issue 1, March 2000.
LODGE, M. and STIRTON, L. Accountability in the Regulatory State. In: BALDWIN, R.,
CAVE, M. and LODGE, M. (Eds). The Oxford Handbook of Regulation, Chapter 15, September
2010.
MATTIUZZO, M. Voto Vencido, Fundamentação Diversa e Fundamentação
Complementar: um estudo sobre deliberação no Supremo Tribunal Federal. 2011. SBDP.
Available at: <http://www.sbdp.org.br/publication/voto-vencido-fundamentacao-diversa-e-
fundamentacao-complementar-um-estudo-sobre-deliberacao-no-supremo-tribunal-federal/>.
MAYER-SCHÖNBERGER, V. and CUKIER, K. Big Data: A Revolution That Will
Transform How We Live, Work, Think. Houghton Mifflin Harcourt, 2013.
MAYER-SCHÖNBERGER, V. and CUKIER, K. The Rise of Big Data: How It's Changing
the Way We Think. Foreign Affairs, vol. 92, no. 3, May/June, 2013. Available at:
<https://www.jstor.org/stable/23526834?seq=1#page_scan_tab_contents>.
MCTIC. Estratégia Brasileira para a Transformação Digital (E-Digital). Brasília, 2018.
Available at: <http://www.mctic.gov.br/mctic/export/sites/institucional/estrategiadigital.pdf>.
MELLO, J. M. P., MENDES, M. and KANCZUK, F. Cadastro Positivo e democratização do
crédito. Folha de São Paulo, March, 2018. Available at:
<https://www1.folha.uol.com.br/opiniao/2018/03/joao-manoel-pinho-de-mello-marcos-
mendes-e-fabio-kanczuk-cadastro-positivo-e-democratizacao-do-credito.shtml>.
MENDES, L. S. Habeas Data e autodeterminação informativa: os dois lados de uma mesma
moeda. In: Centro de Direito, Internet e Sociedade do Instituto Brasiliense de Direito Público
(CEDIS/IDP) (Orgs). Internet & Regulação Saraiva. To be published.
MILLER, A. P. Want Less-Biased Decisions? Use Algorithms. Harvard Business Review.
July 26, 2018. Available at: <https://hbr.org/2018/07/want-less-biased-decisions-use-
algorithms>.
MINISTÉRIO DA TRANSPARÊNCIA, FISCALIZAÇÃO E CONTROLADORIA-GERAL
DA UNIÃO. Aplicação da Lei de Acesso à Informação na Administração Pública Federal.
140
2a. ed. rev., atu. e amp. Brasília. 2016. Available at:
<http://www.acessoainformacao.gov.br/central-de-
conteudo/publicacoes/arquivos/aplicacao_lai_2edicao.pdf>.
NEW, J. and CASTRO, D. How Policymakers can foster algorithmic accountability. Center
for Data Innovation, May 2018. Available at: <http://www2.datainnovation.org/2018-
algorithmic-accountability.pdf>.
NIKLAS, J. SZTANDAR-SZTANDERSKA, K. and SZYMIELEWICZ, K. Profiling the
Unemployed in Poland: social and Political Implications of Algorithmic Decision Making.
Fundacja Panoptykon. Warsaw, 2015.
NISSENBAUM, H. Computing and Accountability. In: COHEN, J. Communications of the
ACM, vol. 37, issue 1, January 1994.
OETER, S. Fundamental Rights and Their Impact on Private Law – Doctrine and Practice
Under the German Constitution. 12 Tel Aviv U. Stud. L. 7, 1994. Available at:
<https://heinonline.org/HOL/LandingPage?handle=hein.journals/telavusl12&div=4&id=&pag
e=>.
O’NEIL, C. Weapons of Math Destruction – How Big Data Increases Inequality and
Threatens Democracy. New York: Crown, 2016.
QUOD. Available at: <https://www.quod.com.br/>.
PANOPTYKON. Available at: <https://en.panoptykon.org/about>.
PARASURAMAN, R. and MILLER, C. A. Trust and etiquette in high-criticality automated
systems. Communications of the ACM – Human-computer etiquette, vol. 47 issue 4, April,
2004. Available at: <https://dl.acm.org/citation.cfm?id=975844>.
PASQUALE, F. The Black Box Society. Harvard University Press. January, 2015.
REISMAN, D., et al. Algorithmic Impact Assessment: A Practical Framework for Public
Agency Accountability. AI Now, April 2018.
RIOS, R. R. and SILVA, R. da. Democracia e direito da antidiscriminação:
interseccionalidade e discriminação múltipla no direito brasileiro. Cienc. Cult., São Paulo,
vol. 69, n. 1, p. 44-49, March, 2017. Available at:
<http://cienciaecultura.bvs.br/scielo.php?script=sci_arttext&pid=S0009-
67252017000100016&lng=en&nrm=iso>.
141
SANDIG, C. et al. An Algorithm Audit. Data and Discrimination: Collected Essays. 2014.
Available at: <http://www-
personal.umich.edu/~csandvig/research/An%20Algorithm%20Audit.pdf>.
SARLET, I. W. Direitos Fundamentais e Direito Privado: algumas considerações em torno
da vinculação dos particulares aos direitos fundamentais. Revista dos Tribunais Online.
Available at: <http://www.direitocontemporaneo.com/wp-content/uploads/2018/03/SARLET-
Direitos-fundamentais-e-direito-privado.pdf>.
SARMENTO, D. Direitos Fundamentais e Relações Privadas. Rio de Janeiro: Lumen Juris,
2004.
SCIENCE AND TECHNOLOGY COMMITTEE of the House of Commons. Algorithms in
decision-making. Fourth Report of Session 2017-19. May 23, 2018. Available at:
<https://publications.parliament.uk/pa/cm201719/cmselect/cmsctech/351/351.pdf>.
SEATTLE INFORMATION TECHNOLOGY. About the Surveillance Ordinance.
September 1, 2017. Available at:
<https://www.seattle.gov/tech/initiatives/privacy/surveillance-technologies/about-surveillance-
ordinance>.
SCHAUER, F. Profiles, Probabilities, and Stereotypes. Belknapp Press. April, 2016.
SHNEIDERMAN, B. Algorithmic Accountability: Designing for Safety. Raddliffe Institute
for Advanced Study Harvard University. March 22, 2018. Available at:
<https://www.radcliffe.harvard.edu/video/algorithmic-accountability-designing-safety-ben-
shneiderman>.
SHNEIDERMAN, B. Opinion: The dangers of faulty, biased, malicious algorithms requires
independent oversight. Proceedings of the National Academy of Sciences of the United States
of America, November 29, 2016
SILVA, V. A. da. A Constitucionalização do Direito: os direitos fundamentais nas relações
entre particulares. São Paulo: Malheiros Editores, 2014.
SILVA, V. A. da. O proporcional e o razoável. Revista dos Tribunais, 2002.
SILVER, D. et al. Mastering the game of Go with deep neural networks and tree search.
Nature, vol. 529, January 28, 2016. Available at: <https://storage.googleapis.com/deepmind-
media/alphago/AlphaGoNaturePaper.pdf>.
SILVER, D. et al. Mastering the game of Go without human knowledge. Nature, vol. 550,
January 19, 2017. Available at:
142
<https://www.nature.com/articles/nature24270.epdf?author_access_token=VJXbVjaSHxFoct
QQ4p2k4tRgN0jAjWel9jnR3ZoTv0PVW4gB86EEpGqTRDtpIz-2rmo8-
KG06gqVobU5NSCFeHILHcVFUeMsbvwS-lxjqQGg98faovwjxeTUgZAUMnRQ>.
SIMON, H. A. Spurious Correlation: A Causal Interpretation. Journal of the American
Statistical Association, vol. 49, September 1954. Available at:
<http://digitalcollections.library.cmu.edu/awweb/awarchive?type=file&item=33513>.
SUPREMO TRIBUNAL FEDERAL. Ministra Carmen Lúcia anuncia início de
funcionamento do Projeto Victor, de inteligência artificial. Notícias STF. August 30, 2018.
Available at: <http://www.stf.jus.br/portal/cms/verNoticiaDetalhe.asp?idConteudo=388443>.
THE COMMITTEE OF EXPERTS ON INTERNET INTERMEDIARIES (MSI-NET) of the
Council of Europe. ALGORITHMS AND HUMAN RIGHTS - Study on the Human Rights
Dimensions of Automated Data Processing Techniques (in particular algorithms) and
Possible Regulatory Implications. Council of Europe. March, 2018. Available at:
<https://rm.coe.int/algorithms-and-human-rights-en-rev/16807956b5>.
THE ECONOMIST. How an algorithm may decide your career. June 21, 2018. Available at:
<https://www.economist.com/business/2018/06/21/how-an-algorithm-may-decide-your-
career>.
THE GUARDIAN. The Cambridge Analytica Files – A year-long investigation into
Facebook, data, and influencing elections in the digital age. Available at:
<https://www.theguardian.com/news/series/cambridge-analytica-files>.
THE NEW YORK CITY COUNCIL. Automated decision systems used by agencies – Law
number 2018/049. January 11, 2018. Available at:
<https://legistar.council.nyc.gov/LegislationDetail.aspx?ID=3137815&GUID=437A6A6D-
62E1-47E2-9C42-461253F9C6D0>.
TRIBUNAL DE JUSTIÇA DO ESTADO DE MINAS GERAIS. TJMG utiliza inteligência
artificial em julgamento virtual. November 07, 2018. Available at:
<https://www.tjmg.jus.br/portal-tjmg/noticias/tjmg-utiliza-inteligencia-artificial-em-
julgamento-virtual.htm#.XC1Vby2ZO1s>.
TUSHNET, M. The issue of state action/horizontal effect in comparative constitutional law.
Oxford University Press and New York University School of Law, vol. 1, number 1, 2003.
VIGEN, Tyler. EBay's Total Gross Merchandise Volume Correlates With Visitors to
Disney Worlds Animal Kingdom. Tylervigen, Spurious Correlations. Available at:
<http://tylervigen.com/view_correlation?id=28571>.
143
VIGEN, T. Spurious correlations. Available at: <http://www.tylervigen.com/spurious-
correlations>.
WALLACE, N. and CASTRO, D. The Impact of the EU’s New Data Protection Regulation
on AI. Center for Data Innovation, March 26, 2018. Available at:
<https://www.datainnovation.org/2018/03/the-impact-of-the-eus-new-data-protection-
regulation-on-ai/>.
WORLD WIDE WEB FOUNDATION. Algorithmic Governance: Applying the Concept to
Different Country Contexts. July, 2017.
ZARSKY, T. Z. Transparent Predictions. University of Illinois Law Review, vol. 2013, no.
4. Available at: <https://www.illinoislawreview.org/wp-content/ilr-
content/articles/2013/4/Zarsky.pdf>.
USA Cases:
Blum v. Yaretsky, 457 U.S. 991 (1982).
Griggs v. Duke Power Co., 401 U.S. 424 (1971).
Marsh v. Alabama, 326 U.S. 501 (1946).
Reitman v. Mulkey, 387 U.S. 369 (1967).
Shelley v. Kraemer, 334 U.S. 1 (1948).
German Cases:
BUNDESVERFASSUNGSGERICHT. Decision in “Stadium Ban” proceedings clarifies the
indirect horizontal effects of the right to equality in private law relations. Press Release No.
29/2018, April 27, 2018. Available at:
<https://www.bundesverfassungsgericht.de/SharedDocs/Pressemitteilungen/EN/2018/bvg18-
029.html>.
BUNDESVERFASSUNGSGERICHT. Headnotes to the Order of the Firste Senate of 11
April 2018 – 1 BvR 3080/09. Available at:
144
<https://www.bundesverfassungsgericht.de/SharedDocs/Entscheidungen/EN/2018/04/rs20180
411_1bvr308009en.html;jsessionid=34F6550A81C53BC9257433E87D7CF087.1_cid393>.
BVerfGE, 7, 198.
BVerfGE 65,1, Volkszählung.
Brazilian Cases:
5ª PROMOTORIA DE JUSTIÇA DE TUTELA COLETIVA DE DEFESA DO
CONSUMIDOR E DO CONTRIBUINTE DA CAPITAL. Petição inicial, Inquérito Civil n.
347/5ª PJDC/2016.
MINISTÉRIO DA JUSTIÇA. Nota Técnica n. 92. 2018, §39. Available at:
<http://www.cmlagoasanta.mg.gov.br/abrir_arquivo.aspx/PRATICAS_ABUSIVAS_DECOL
ARCOM?cdLocal=2&arquivo=%7BBCA8E2AD-DBCA-866A-C8AA-
BDC2BDEC3DAD%7D.pdf>.
STF. Recurso Extraordinário 201.819-8, Rio de Janeiro, 2005. Available at:
<http://redir.stf.jus.br/paginadorpub/paginador.jsp?docTP=AC&docID=388784>.
Legislation:
BRAZIL. Constitution. 1988. Available at:
<http://www.planalto.gov.br/ccivil_03/Constituicao/Constituicao.htm>.
BRAZIL. Bill n. 7.716. January 05, 1989. Available at:
<http://www.planalto.gov.br/ccivil_03/LEIS/L7716.htm>.
BRAZIL. Bill n. 8.078 (CDC). September 11, 1990. Available at:
<http://www.planalto.gov.br/ccivil_03/Leis/L8078compilado.htm>.
BRAZIL. Bill n. 9.029. April 13, 1995. Available at:
<http://www.planalto.gov.br/CCIVIL_03/LEIS/L9029.HTM>.
BRAZIL. Bill n. 10.406 (Civil Code). January 10, 2002. Available at:
<http://www.planalto.gov.br/ccivil_03/leis/2002/l10406.htm>.
145
BRAZIL. Bill n. 12.288.. July 20, 2010. Available at:
<http://www.planalto.gov.br/ccivil_03/_Ato2007-2010/2010/Lei/L12288.htm>.
BRAZIL. Bill n. 12.414. June 09, 2011. Available at:
<http://www.planalto.gov.br/ccivil_03/_Ato2011-2014/2011/Lei/L12414.htm>.
BRAZIL. Bill n. 12.527 (LAI). November 18, 2011. Available at:
<http://www.planalto.gov.br/ccivil_03/_ato2011-2014/2011/lei/l12527.htm>.
BRAZIL. Bill n. 12.711. August 29, 2012. Available at:
<http://www.planalto.gov.br/ccivil_03/_ato2011-2014/2012/lei/l12711.htm>.
BRAZIL. Bill n. 12.965, April, 23, 2014. Available at:
<http://www.planalto.gov.br/ccivil_03/_ato2011-2014/2014/lei/l12965.htm>.
BRAZIL. Bill n. 13.709 (LGPD). August 14, 2018. Available at:
<http://www.planalto.gov.br/ccivil_03/_Ato2015-2018/2018/Lei/L13709.htm>.
BRAZIL. Draft Bill n. 441. 2017. Available at:
<http://www.camara.gov.br/proposicoesWeb/fichadetramitacao?idProposicao=2160860>.
BRAZIL. Decree n. 7.724. May 16, 2012. Available at:
<http://www.planalto.gov.br/ccivil_03/_ato2011-2014/2012/Decreto/D7724.htm>.
BRAZIL. Decree n. 8.771. May 11, 2016. Available at:
<http://www.planalto.gov.br/ccivil_03/_Ato2015-2018/2016/Decreto/D8771.htm>.
BRAZIL. Executive Order n. 869. December 27, 2018. Available at:
<http://www.planalto.gov.br/ccivil_03/_Ato2015-2018/2018/Lei/L13709.htm>.
EUROPEAN UNION. Regulation (EU) 2016/679 (General Data Protection Regulation).
May 25, 2018. Available at: <https://gdpr-info.eu/>.
POLAND. The Constitution of the Republic of Poland. April 02, 1997. Available at:
<https://www.sejm.gov.pl/prawo/konst/angielski/kon1.htm>.